为了解决分布式负荷响应调频指令中的效率问题，提出了一种基于强化学习的负荷聚合商电价激励响应调频创新策略，在该策略中，构建聚合商和负荷集群博弈模型，聚合商根据调频指令和激励电价策略调整激励电价，而负荷根据自身用电成本调节用电功率，灵活地响应调频指令，采用多智能体软演员批评家（multi-agent soft actor-critic, MASAC）算法求解。结果表明电价激励方法可以使得负荷有效响应调频指令，通过MASAC算法不仅可以优化决策过程，还能有效降低运算复杂性，实现高效的动态调节。可见，该方法为电力系统的频率调节提供了一种有效的解决方案，具有重要的理论意义和实际应用价值。

In order to solve the efficiency issues in distributed load responses to frequency regulation commands, an innovative strategy was introduced based on reinforcement learning for load aggregators' pricing incentives in response to frequency commands. Within this strategy, a game-theoretic model between the load aggregators and load clusters was constructed, and the load aggregators adjust incentive prices based on frequency commands and their pricing strategies, while loads adjust their power consumption based on their own electricity costs to flexibly respond to the frequency commands. The multi-agent soft actor-critic (MASAC) algorithm was used to investigate the solution. The results show that the pricing incentive method enables effective load response to frequency commands, and the use of the MASAC algorithm not only optimizes the decision-making process but also significantly reduces computational complexity, achieving efficient dynamic adjustment. It is concluded that this method provides an effective solution for frequency regulation in power systems, offering significant theoretical significance and practical value.